Semantic Segmentation and Domain Adaptation of Aerial Images Using Binary Space Partitioning

Daniel Gritzner

doi:10.51202/9783186890108

Search publication

Bibliographic data

Edition: 1/2026
ISBN-Print: 978-3-18-589010-9
ISBN-Online: 978-3-18-689010-8
Publisher: VDI Verlag, Düsseldorf
Series: Informatik/ Kommunikation
Volume: 890
Language: English
Pages: 146
Product Type: Monograph

1. Vorwort No access
2. Contents No access
3. Acronyms No access
4. Notation No access
5. Abstract No access
6. Kurzfassung No access
1. 1.1 Motivation No access
2. 1.2 Structure No access
1. 2.1 Semantic Segmentation No access
2. 2.2 Domain Adaptation No access
1. 1. 3.1.1 Aerial Images No access
  2. 3.1.2 Segmentation No access
  3. 3.1.3 Semantic Segmentation Evaluation Metrics No access
  4. 3.1.4 Fourier Transform No access
2. 1. 3.2.1 Artificial Neural Networks No access
  2. 3.2.2 Whitening No access
  3. 3.2.3 Activation Functions No access
  4. 3.2.4 Normalization No access
  5. 3.2.5 Loss Functions No access
  6. 3.2.6 Backpropagation No access
  7. 3.2.7 Optimizers No access
  8. 3.2.8 Generalization No access
  9. 3.2.9 Convolutional Neural Networks No access
  10. 3.2.10 Ensemble Methods No access
3. 1. 3.3.1 Transfer Learning No access
  2. 3.3.2 Domain-Adversarial Neural Networks No access
  3. 3.3.3 Cycle-GAN No access
  4. 3.3.4 Self-Training No access
4. 1. 3.4.1 Binary Space Partitioning No access
  2. 3.4.2 Signed Distance Functions No access
  3. 3.4.3 Quadtrees No access
  4. 3.4.4 k-dTrees No access
1. 4.1 Model Architecture No access
2. 4.2 Differentiable BSP Tree Rendering No access
3. 4.3 Shape-optimizing Loss Function No access
4. 1. 4.4.1 BSP Trees as Ground Truth No access
  2. 4.4.2 Ablation No access
  3. 4.4.3 Encoder/Backbone No access
  4. 4.4.4 Comparison to Other Segmentation Models No access
  5. 4.4.5 Region Map Analysis No access
1. 1. 5.1.1 Patch Extraction No access
  2. 5.1.2 Learn Mean and Covariance Matrices No access
  3. 5.1.3 Compute Whitening Matrices No access
  4. 5.1.4 Transfer Appearance No access
  5. 5.1.5 Combine Transferred Patches No access
  6. 5.1.6 Adaptation Direction No access
2. 5.2 Feature Adaptation No access
3. 5.3 Self-Training No access
4. 1. 5.4.1 Appearance Adaptation No access
  2. 5.4.2 Full Domain Adaptation No access
1. 6.1 Datasets No access
2. 6.2 Testing Procedure No access
3. 6.3 Semantic Segmentation No access
4. 1. 6.4.1 Comparison to State of the Art No access
1. 7.1 Summary No access
2. 7.2 Future Work No access
Appendix A No access Pages 103 - 106
Bibliography No access Pages 107 - 120
Publications/Author No access Pages 121 - 128

Semantic Segmentation and Domain Adaptation of Aerial Images Using Binary Space Partitioning

Search publication

Bibliographic data

Table of contents